[petsc-dev] Improving nightly builds for maint

[Lisandro Dalcin]

On 20 October 2017 at 21:12, Jed Brown <jed at jedbrown.org> wrote:
> Barry Smith <bsmith at mcs.anl.gov> writes:
>
>>> On Oct 18, 2017, at 3:10 PM, Lisandro Dalcin <dalcinl at gmail.com> wrote:
>>>
>>> On 18 October 2017 at 22:36, Satish Balay <balay at mcs.anl.gov> wrote:
>>>> On Wed, 18 Oct 2017, Lisandro Dalcin wrote:
>>>>
>>> No, please don't!  Be green, Satish!
>>>
>>> What's the point of running a 3D problem under valgrind for 2 hours?
>>
>>    Find memory corruption bugs that won't be found otherwise?
>
> So use three elements and two time steps.
>

Indeed. That's what I meant. This particular test I'm talking about
runs a converge test under grid refinement for a 3D problem. Lots of
builds in maint are still yellow just because this test is not able to
finish in 600 seconds in non-valgrind builds!

>
>>    But note that the original test problems should be taking only a short time so even with valgrind should not take 2 hours.
>

Satish set TIMEOUT to more than 2 hours. If a test ever fails because
of a deadlock, the build worker will be stuck for 2 hours. Of course,
we will likely notice, but still...



-- 
Lisandro Dalcin
============
Research Scientist
Computer, Electrical and Mathematical Sciences & Engineering (CEMSE)
Extreme Computing Research Center (ECRC)
King Abdullah University of Science and Technology (KAUST)
http://ecrc.kaust.edu.sa/

4700 King Abdullah University of Science and Technology
al-Khawarizmi Bldg (Bldg 1), Office # 0109
Thuwal 23955-6900, Kingdom of Saudi Arabia
http://www.kaust.edu.sa

Office Phone: +966 12 808-0459